The live video technology has been widely used and the TV belongs to such technology. The live video technology is characterized by the timeliness, that is, the live video content can not be viewed if the time is missed. For a client who wants to view the live video content but his/her time can not be guaranteed, he/she either waits for the re-broadcast of the video content or views the video content by the way of on demand. However, it is uncertain whether the live video content will be re-broadcasted, whether the live video content will become the video content on demand and when the video content on demand is put on the line. For satisfying the clients, the technology of storing the live video content appears. Such storage may be realized on a server or by a storage device at the client.
The realization at the client refers to that a storage device is configured at the client and the client sets information of channels to be stored, the start time and the end time of the storage or the like, on the basis of his/her favor. This method has some disadvantages in that the capacity of the storage device at the client is limited, thus only a certain amount of video content can be stored, and generally, the storage device stores only the live video content for one channel at each time period. If the client needs to store the live video content for multiple channels, or needs to store the live video content for a long time period, the storage device at the client can not satisfy this requirement. Additionally, the most important problem of this method lies in that the client is required to set recording channel and time for a program in advance according to a program schedule. For an event that happens unexpectedly without a program schedule, the storage information can not be set and the storage can not be performed because the client does not know the start time of the live video content.
The realization at the server end refers to that a storage device is configured on the server for storing live video content for all channels. If the user needs the live video content of a channel prior to the current time, the client sends to the server a request including channel information and information of the start time of the live video data, for example, the current time is 12:00, Dec. 15, 2009, and a client requests viewing of the live video data of channel A with a start time is 11:00:35, Dec. 1, 2009, the server searches for the storage position of the live video content of channel A with a start time is 11:00:35, Dec. 1, 2009 on receiving the request from the client, and obtains the live video data from the storage position of the live video content and sends the live video data to the client. It seems like the live video content after time-shift is viewed by the client, which is referred to as live time-shift technology.
Because the storage capability of the server is very high, that is, the server can store all live video content for all live channels, which can meet the requirement of clients to view the live video content that is not specific.
To distinguish from the live video, the live video after time-shift is referred to as live time-shift video hereinafter.
The live time-shift technology using the server for storage has been relatively mature, a Chinese patent (publication number: CN 101207795A, abbreviated as reference patent 1) titled “live server with time-shift function and implementing method thereof” discloses a representative live time-shift technology using a server for storage. In this patent, the live server is modified to add a storage function, thus the live server can also be used as a time-shift server. During the storage, upon receiving live video sources, the modified live server issues live video data while storing live video data, and adds video attributes and time information to the stored live video data. In this way, live time-shift video service can be provided to clients according to the index of channel and time information.
In published documents, the live time-shift technology (including the technology of the reference patent 1) performs the transmission of live video content in the manner of unicast. The unicast means that a data channel is established between each client and the server, thus the number of occupation of server resources is linearly correlated with the number of the clients. If the number of the clients increases sharply, the server resources increase linearly, thus it is going against mass application of the unicast.
At present, the P2P technology is widely used in the video transmission service, a Chinese patent (publication number: CN 101075954A, abbreviated as reference patent 2) titled “point-to-point streaming media communication system and method for configuring resources in super node thereof” is a much representative P2P video transmission technology. In the reference patent 2, a method for realizing streaming media transmission system based on the P2P technology is described. Such technology may be used in an on-demand service and a live service.
In published documents about the P2P live technology, the value of time length of the memory buffer in the server is typically in the order of second (generally within 150 s). Because the length of the memory buffer affects the time delay for displaying the live video on the client terminal, the increasing of the time delay will affect the visual effect of the client.